Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 13

Question 120

An organization publishes MCP tools for searching internal technical documentation. A semantic-search tool returns ranked snippets that may not be complete documents. What is the most defensible next step?

  1. Describe deterministic lookup versus semantic search use cases explicitly
  2. Document optional fields, cursor behavior, and what confidence actually represents
  3. Document that search outputs are excerpts with source references and may need follow-up reads
  4. Validate and constrain readable paths in the tool implementation
  5. Specify its corpus, query shape, and result guarantees in the description

Explanation

Correct Answer

C:

An explicit limitation discourages Claude from treating a truncated hit as authoritative coverage of a whole document.

Incorrect Answers

A:

A direct identifier lookup gives precise retrieval when a key is known; embedding search is better for conceptual matching, not exact ID verification; it protects against the case where two tools share a capability, but one is optimized for exact identifier lookups, rather than the failure described here: a semantic-search tool returns ranked snippets that may not be complete documents.

B:

Typed, well-described results let the agent distinguish absent values from negative findings and paginate without assumptions; that applies when a results object has optional confidence and pagination fields, whereas this case requires a response to a semantic-search tool returns ranked snippets that may not be complete documents.

D:

Descriptions help selection but do not provide security; a server-side allowlist must protect unauthorized files; this is useful for a file-read tool accepts any path string and can access outside the workspace, but it leaves the stated requirement to state the limitations of retrieval results unmet.

E:

This measure fits a different trigger: a tool named query has no indication of which index it searches. In that setting, Model selection depends on clear capability metadata; an opaque name leaves the assistant guessing which source will answer the user's question; it does not resolve the present need to state the limitations of retrieval results.

Question 121

A developer uses Claude Code to change a risky authentication module. A typo in a known Markdown file is approved for immediate correction. Which approach satisfies the stated constraint?

  1. Apply a direct targeted edit and verify the diff
  2. Use read-only search and inspection to map the responsible files
  3. Present the plan and wait for authorization before enabling write operations
  4. Use plan mode to inspect the codebase and propose steps without applying edits

Explanation

Correct Answer

A:

A small, well-scoped low-risk change can be executed directly when the user has approved it. The relevant requirement is to avoid an unnecessary planning cycle, and this measure operates at that decision point.

Incorrect Answers

B:

An exploratory pass establishes architecture evidence; jumping into edits risks changing the wrong component; this is useful for the team does not know where authorization is implemented, but it leaves the stated requirement to avoid an unnecessary planning cycle unmet.

C:

A design review requirement is not satisfied merely by generating a plan if the next stage can still commit changes autonomously; that applies when a proposed refactor requires an explicit human go/no-go, whereas this case requires a response to a typo in a known Markdown file is approved for immediate correction.

D:

This measure fits a different trigger: the engineer wants a proposed design and affected-file list before any modifications. In that setting, Read-only planning supports review of dependencies and risk before the team authorizes source changes; it does not resolve the present need to avoid an unnecessary planning cycle.

Question 122

A Claude Agent SDK workflow delegates repository investigations to subagents. A specialist's large file-search transcript is not useful to the parent. Which mechanism should be selected?

  1. Launch the independent module inquiries in parallel and merge results later
  2. Run the investigation in a subagent and return a concise evidence-based summary
  3. Give the Task-oriented specialist a concrete goal, scope, and expected handoff
  4. Have the worker report changed or inspected paths and evidence for its conclusion

Explanation

Correct Answer

B:

An isolated worker context avoids flooding the coordinator with irrelevant file reads while preserving the actionable findings.

Incorrect Answers

A:

Concurrent subagents can reduce wall-clock time when their files and decisions do not have data dependencies; this is useful for two independent test modules can be investigated separately, but it leaves the stated requirement to isolate exploratory context from the coordinator unmet.

C:

Structured delegated tasks clarify which worker owns the operation and how its outcome will return to the parent; that applies when a parent agent needs a specialist to take a bounded action rather than just offer advice, whereas this case requires a response to a specialist's large file-search transcript is not useful to the parent.

D:

The parent can validate a specialist's claims only when the output identifies what was examined or modified; it protects against the case where a subagent returns a broad recommendation without citing inspected files, rather than the failure described here: a specialist's large file-search transcript is not useful to the parent.

Question 123

An AI research assistant prepares high-stakes organizational recommendations. The cited facts support two materially different recommendations. What should be implemented at the appropriate boundary?

  1. Execute the delegated decision inside the defined thresholds with an audit trail
  2. Present competing interpretations and escalate the trade-off to the responsible decision maker
  3. Apply documented style rules automatically and continue
  4. Mark the affected conclusion as unverified and request further review

Explanation

Correct Answer

B:

When evidence does not uniquely determine action, the architect should avoid manufacturing unwarranted certainty. Here it serves the explicit need to preserve decision uncertainty without adding a separate unverified inference.

Incorrect Answers

A:

Escalation should follow actual authority boundaries; needless handoffs for approved routine cases slow the system without improving safety; it protects against the case where a legal owner has explicitly delegated a narrow approved judgment, rather than the failure described here: the cited facts support two materially different recommendations.

C:

Human review capacity is best reserved for consequential ambiguity; deterministic cosmetic corrections can be handled within the workflow; that applies when a minor stylistic issue does not affect factual or safety requirements, whereas this case requires a response to the cited facts support two materially different recommendations.

D:

This measure fits a different trigger: a key source is inaccessible and no equivalent evidence is available. In that setting, The absence of a crucial source limits what the assistant can responsibly conclude; a plausible summary cannot replace verification; it does not resolve the present need to preserve decision uncertainty.

Question 124

Claude Code is used to modernize a legacy service under production constraints. The task is to locate files referencing an old environment flag. Which option best meets the operational need?

  1. Generate alternative plans with trade-offs and identify missing evidence
  2. Edit the identified key and run the acceptance test
  3. Search for the flag and inspect references using read-only tools
  4. Limit writes to authorized paths and escalate the out-of-scope changes

Explanation

Correct Answer

C:

Locating references is a discovery task; executable patches are premature until intended behavior is understood. For the stated case, the check is whether the system can perform focused exploration without editing; this approach supplies that control.

Incorrect Answers

A:

A decision stage should expose risks and open questions rather than let the assistant silently pick an unreviewed migration strategy; that applies when the team disagrees about the safest migration approach, whereas this case requires a response to the task is to locate files referencing an old environment flag.

B:

A concrete low-risk request with clear verification can move directly to execution without an elaborate architecture phase; this is useful for a one-line configuration adjustment has an approved target and acceptance check, but it leaves the stated requirement to perform focused exploration without editing unmet.

D:

Approval for a plan does not grant unrestricted filesystem rights; actual tool permissions must still match the authorized work; it protects against the case where the final plan is accepted, but some files lie outside the permitted workspace, rather than the failure described here: the task is to locate files referencing an old environment flag.

Question 125

A codebase migration process must decide when to spawn specialists. A database migration depends on a schema inventory that does not yet exist. What is the most appropriate design choice?

  1. Run schema discovery before generating or applying migration steps
  2. Use one owning reviewer or coordinator for the business-rule judgment
  3. Serialize updates or give workers isolated copies before concurrency
  4. Delegate each service to a scoped worker and compare results afterward
  5. Create smaller file- or component-scoped subtasks with explicit output summaries

Explanation

Correct Answer

A:

A migration plan based on guessed schema can break data, so dependency evidence must be present first. The relevant requirement is to make inventory a hard prerequisite, and this measure operates at that decision point.

Incorrect Answers

B:

Independent conflicting interpretations should not each drive a code change; central adjudication preserves consistency; this is useful for a change requires one authoritative interpretation of business rules, but it leaves the stated requirement to make inventory a hard prerequisite unmet.

C:

Parallel actions that change the same state can race; independence must include write effects, not merely different prompts; it protects against the case where several workers share a mutable global configuration during analysis, rather than the failure described here: a database migration depends on a schema inventory that does not yet exist.

D:

This measure fits a different trigger: two services have unrelated APIs and can be analyzed separately. In that setting, Independent boundaries permit concurrency without creating cross-service state conflicts; it does not resolve the present need to make inventory a hard prerequisite.

E:

Artifact-based partitions control context size while leaving traceable, mergeable findings; that applies when a large task repeatedly overwhelms a worker's context window, whereas this case requires a response to a database migration depends on a schema inventory that does not yet exist.

Question 126

A code-inspection MCP server reports results and operational errors to Claude Code. A search query is valid and returns no matches in the target directory. Which response is most reliable?

  1. Return a successful empty hit collection with the applied search scope
  2. Reject incomplete tool results before marking the inspection complete
  3. Return a not-found result with the indexed revision context
  4. Report access denial with path scope but without privileged file contents

Explanation

Correct Answer

A:

Zero results is useful evidence about a bounded search; presenting it as a tool failure would trigger unnecessary recovery.

Incorrect Answers

B:

Structural success is insufficient when required provenance fields are missing; validation must check the actual promised result schema; it protects against the case where a tool reports success but omits a required file hash, rather than the failure described here: a search query is valid and returns no matches in the target directory.

C:

The agent needs to know search and read views differ, so it can refresh or requery rather than citing nonexistent content; this is useful for a path points to a file deleted after the search index was built, but it leaves the stated requirement to represent the absence of matches faithfully unmet.

D:

This measure fits a different trigger: the source file exists but cannot be read due to repository permissions. In that setting, A permission-denied file is not equivalent to missing code; the operator may need an approved access adjustment; it does not resolve the present need to represent the absence of matches faithfully.

Question 127

A research writing assistant must deliver claims that can be independently checked. The task asks for a concise brief but never sets a length or section requirement. What should the team do to resolve this particular gap?

  1. Require effective dates and labels for superseded rules
  2. Define target sections, approximate length, and mandatory findings
  3. Require verification of critical numbers before publishing, even if additional review is needed
  4. Require numeric support or qualified sourcing for comparative claims

Explanation

Correct Answer

B:

A measurable structure improves consistency without requiring every word to follow a rigid template. Here it serves the explicit need to specify observable document structure without adding a separate unverified inference.

Incorrect Answers

A:

Time qualifiers are part of factual correctness when documents describe changing systems; bare claims can mislead; that applies when a report must distinguish current policies from historical descriptions, whereas this case requires a response to the task asks for a concise brief but never sets a length or section requirement.

C:

A prompt can express the quality trade-off explicitly; rushing a confident result is inappropriate when a wrong figure has high cost; it protects against the case where two stakeholders prioritize accuracy over speed for high-impact figures, rather than the failure described here: the task asks for a concise brief but never sets a length or section requirement.

D:

This measure fits a different trigger: a generated comparison uses qualitative adjectives with no supporting figures. In that setting, A clear evaluation rule distinguishes evidence-based comparison from untestable promotional language; it does not resolve the present need to specify observable document structure.

Question 128

An agent runs a bounded support workflow with multiple optional tool calls. A user cancels a request while a queued tool operation has not started. Which control is most directly relevant?

  1. Check cancellation state before dispatching queued tool work
  2. Enforce a maximum tool-call or iteration budget and a fallback outcome
  3. Use rate-aware concurrency limits and jittered backoff
  4. Log authorized tool names, request IDs, outcomes, and decision context

Explanation

Correct Answer

A:

A cancellation flag in application state can block new operations; telling the model to stop is not sufficient for queued execution.

Incorrect Answers

B:

This measure fits a different trigger: a tool-driven conversation keeps calling the same read-only endpoint without progress. In that setting, Without a host-side bound, repeatedly proposing the same action can consume resources without resolving the customer's issue; it does not resolve the present need to respect cancellation before further side effects.

C:

Application-side scheduling controls aggregate demand; independent immediate retries can amplify throttling across sessions; this is useful for a backend becomes rate limited while many agents share it, but it leaves the stated requirement to respect cancellation before further side effects unmet.

D:

Structured execution logs establish which external action occurred and on what evidence without assuming that model prose is a reliable audit record; that applies when support needs a record of why an agent issued each sensitive call, whereas this case requires a response to a user cancels a request while a queued tool operation has not started.

Question 129

A release team operates Claude Code as part of automated change analysis. An agent proposes a code change that CI cannot reproduce in a fresh checkout. How should the workflow be adjusted?

  1. Compare findings against adjudicated examples and track precision and recall
  2. Re-run build and tests from the committed revision in a clean environment
  3. Emit a fixed verdict schema and reject unparseable output
  4. Block automatic merge and provide evidence for maintainer review

Explanation

Correct Answer

B:

Verification from a clean checkout detects hidden local state and makes results reproducible across runners. For the stated case, the check is whether the system can verify artifacts against an immutable source revision; this approach supplies that control.

Incorrect Answers

A:

CI model output can drift; longitudinal labeled evaluation measures whether the review remains useful rather than merely verbose; that applies when the model-generated review itself must be measured for false positives over time, whereas this case requires a response to an agent proposes a code change that CI cannot reproduce in a fresh checkout.

C:

This measure fits a different trigger: the pipeline prints a detailed narrative but downstream jobs need a verdict field. In that setting, A deterministic parser must be able to identify pass, needs-review, and blocked states without interpreting conversational language; it does not resolve the present need to verify artifacts against an immutable source revision.

D:

A high-risk finding should not be auto-dismissed by the same generative process that produced the report; this is useful for a reviewer finds a high-risk issue requiring a person's decision, but it leaves the stated requirement to verify artifacts against an immutable source revision unmet.

Question 130

A code assistant integrates MCP tools maintained by different engineering groups. A deployment tool accepts an environment name with many misspelled variants. Which implementation decision is the best fit?

  1. Use an enumerated argument value set and validate it on the server
  2. Version breaking schema changes or provide a backwards-compatible migration
  3. Return an empty success result for legitimate no-match outcomes
  4. Ignore instructions in tool metadata that conflict with higher-priority security policy
  5. Redact secrets before returning tool content to Claude

Explanation

Correct Answer

A:

Schema constraints and server checks prevent an unintended target resulting from free-form environment strings. For the stated case, the check is whether the system can limit arguments to supported environments; this approach supplies that control.

Incorrect Answers

B:

Unexpected field renames can break tool consumers and downstream validators even when a tool's description still sounds correct; it protects against the case where an internal tool is versioned while several clients depend on its outputs, rather than the failure described here: a deployment tool accepts an environment name with many misspelled variants.

C:

No matching build is a valid result distinct from an operational error; mixing them causes unnecessary retries and escalation; this is useful for a build lookup returns no match and callers mistake it for a crashed server, but it leaves the stated requirement to limit arguments to supported environments unmet.

D:

This measure fits a different trigger: one tool's description encourages bypassing repository review rules. In that setting, Descriptions are useful for discovery, not an authority source; a malicious tool can attempt prompt injection through its advertised text; it does not resolve the present need to limit arguments to supported environments.

E:

Tool responses become model context and may be quoted later; a restricted data boundary must filter credential material; that applies when an API key appears in an MCP tool's result object, whereas this case requires a response to a deployment tool accepts an environment name with many misspelled variants.

Claude Certified Architect - Foundations (CCAR-F) (CCA-F) Practice Tests